GPS outputting magnetic variation?

Hi All,

I am working on an application where I need to know the magnetic variation
at a certain location. According to the NMEA specs, the RMC sentence has a
field for this. My GPS however, a u-Blox module (SiRF), does not output the
variation. So I am wondering: is this info sent to a GPS in the almanac data
or is this field only present for 'other purposes'.

Meindert

 

The GPS as a system doesn't care at all for any north orientation
(except for the ECEF inherent north orientation). Almanach and ephemeris
contain only data relevant to time and satellite position and broadcast.
As the satellites know nothing about the receiver position, they can not
transmit any information related to the GPSR position.

The receiver has to maintain a table for magnetic variation and apply
this data based on his location, and probably do some interpolation.
Recently someone posted a link to an internet site with
information/tools/data sets to do this.

This is quite similiar to geoid correction that some (most) receivers do
- models and/or tables that have to be applied based on the current
position. A classical memory size vs. computational cost vs. accuracy
triangle constraint.

Actually very few GPS sensors/modules seem to have an option to define
TrueNorth vs. MagneticNorth. If they have, they will probably also
output magnetic deviation in RMC.

Your only choice is to apply the magnetic deviation model yourself or
use a different receiver board. I don't know currently which
board/chipset/firmware supplies magnetic deviation, but I could browse a
few datasheets. I would think that at least the Garmin
trackpacks/engines do it.

Even if some devices do - it is probably important for you to know HOW
GOOD they do ...

Im wondering if you could check this by issuing a receiver coldstart
with supplied position init. Maybe some receivers output magnetic
variation based on the initialized position value for a while, even if
it is not confirmed by a position fix. That way you could send a grid
series of inits to the receiver, read out the magnetic deviation value
and compare them with the more accurate tables/models on the internet.


Interesting problem ;-)




- Carsten

 

Hmm, I just found this in a SIRFIIe document dating from October 2001:

---
2. SiRF Technology Inc. does not support magnetic declination. All
“course over ground” data are geodetic WGS84
directions.
---

They may still add it later, but I wouldn't count on it. This feature is
not important for the current target markets of SIRF receivers.

- Carsten

 

I know, but I can imagine the in the almanac data, the position of the
magnetic north is included and that the GPS can calculate the variation for
the position it is on.

Many GPSR's can be set to output true or magnetic heading. At least the
marine types can. But some of them even allow the user to input the magnetic
variation. So I'm still not sure if it is a precise figure or just a wild
guess.

That would be interesting yes, given the fact that I am trying to create a
low cost solution that should match a gyro-compass as close as possible.

It sure is

Meindert

 

That's exactly what I read in my SiRF document :-(

Meindert

 

And in case you try - it is also true for the TIM-LP/Antaris chipset.

- Carsten

 

I think Meindert was suggesting that the position of the magnetic north
pole (which is constantly shifting) could be provided. Unfortunately that
wouldn't be sufficient to determine the local magnetic deviation since the
magnetic field lines do not generally point straight at the magnetic north
pole. There are lots of local variations so the GPS needs a lookup table
and interpolation to determine the deviation at a given location.

 

Well, the position of the magnetic north is *not* well know. That is, it
varies over time. I thought that maybe, a GPS can calculate the average
variation based on it's position and the position of the magnetic north. But
the variation is also dependent on other factors, I suppose. So it would
need some sort of grid, with variations for the grid regions throughout the
world.

Yes. But on certain ships, it is required to have a true north device, and a
'normal' GPS does not qualify for that, since it only can give a bearing
when moving. The solution of course is a gyro, but what I am trying to
achive is to use a fluxgate (magnetic) and use the variation, possibly from
the GPS, to turn the magnetic course into the true course.

Meindert

 

Just because it varies position doesn't mean it isn't well known:
http://www.geolab.nrcan.gc.ca/geomag/northpole_e.shtml

 

Hmm. I guess all simple MD calculations in GPSRs have larger errors than
those caused by the shifting magnetic north?

How much does it typically shift, and how fast? Is it a regular pattern,
like the earths precession?

Hmm, can answer that myself:

http://www.geolab.nrcan.gc.ca/geomag/long_mvt_nmp_e.shtml

Seems to be 40km per year currently. Wow! But it seems to be a fairly
straight line so far.

Do GPSRs with magnetic north capability take this into account when
calculating magnetic north?

Well certainly, GPS does not transmit the current position of magnetic
north. And probably no GNSS currently being planned will do that. They
could as well transmit the position of the nearest bakery. Well, not
quite ...

It seems that a separate data connection would be neeeded to receive
uptodate information. Seems that data is usually available with a delay
of one day. If I understand Meindert's application well, he would be
interested in something like a realtime transmission. Seems that this is
not possible.

Of course, using very precise GPS attitude/gyro systems and a very
precise compass, + a huge datatable, the position of magnetic north
could be computed right on place ;-)



- Carsten

 

I don't know how accurate this has to be to be useful for your
application. The long term variation seems to be a straight line.
Although it seems to 'circle' over short time periods.

You mean the problem is to have TrueNorth on an idling platform? Well, I
guess on a ship you can easily use two receivers/antennas spaced as far
as possible. Could as well serve as a backup receiver.

http://www.sanav.com/gps_compass/sk-500.htm

It will also allow you to determine ship yaw accross the ships main
axis. (forgive - I'm not a mariner)

- Carsten

 

[This may be a duplicate post.
When I tried to SEND the first time
things hung up and never went to "posted."]

==========================

There are models with software code
that can be used to calculate the
magnetic variation (a.k.a. magnetic declination)
for a given lat/long with a target accuracy
of 1 degree for a given date, not including
local magnetic anomalies like big mountains
of iron ore. The model parameters are updated
every 5 years. The tables in GPS receivers
are likely derived from these models.

From Magnetic FAQ
http://www.ngdc.noaa.gov/seg/potfld/faqgeom.shtml#q1b

" 6d. How do I get the latest [world magnetic] model?

NGDC/WDC for Solid Earth Geophysics archives and distributes
the IGRF, WMM and other models.
You can download the software and latest model available
at no charge or order the software, model
and documentation package for a slight fee. Go to

http://www.ngdc.noaa.gov/seg/potfld/magmodel.shtml. "

The World Magnetic Model (WMM) is included in my freeware
great circle/ellipsoid program GCGC* to calculate
magnetic bearings from true bearings.

Cheers, 73,

Ron McConnell



WGS-84: N 40º 46' 57.9" W 74º 41' 21.9"
Magnetic Variation/Declination = 13º in early 2004

FN20ps77GV75 per w2iol
or FN20ps77GU46 per K2RIW

* GCGC excutable and source code at
http://home.earthlink.net/~rcmcc

 

I know that solution. But still, at about $5000-7000 quite expensive. Two
normal GPS receivers will indeed do it, provider they are far enough apart.

Meindert

 

Must be. How would they otherwise be able to do that?

Ok, but then there must be something transmitted to the GPS to determine the
variation. I cannot imagine that this is based on a fixed table inside the
GPS, being useless after a few years.

Meindert

 

Some systems even use just two switchable antennas. But usually nowadays
it's cheaper to use two GPS modules with a long run of RS232/422 cable
than to use costly low-loss antenna cable ;-)


- Carsten

 

Well they could just assume a fixed position and be satiefied with the
accuracy they can achieve that way.

No, there is not. The magnetic course correction in consumer GPSRs is
not intended to be too accurate. I guess it's just accurate enough to
compensate for a decent conventional hiking compass.
Why do you think some of the better receivers have this option to input
MD manually?

Besides that, KNOWING the magnetic north is only one part of the job.
The magnetic declination is not equally distributed over the globe but
has non-linearities that can only be expressed in tables. There is not
enough bandwith in those 50bps of GPS to transmit declination data, not
even as a corse grid.

I would be surprised if GNSS designers think about magnetic declination
or magnetic north. Their work is about getting rid of these things.
WGS84 doesn't care about magnetic north.

You can, of course browse the GPS space segment documents to find out
what IS broadcast and what not. Believe me, it's not very much indeed,
and they already used some smart methods to shrink the necessary data in
this low bandwith signal. The signal spec is (and has to be) still valid
from the seventies!

- Carsten

 

It is certainly done with tables and algorithms inside the gps if it is
done at all. There are less useful after a few years but that is what
firmware upgrades can fix. There is nothing transmitted from the satellites.

Dale

 

I found the answer on some forum, coming from Garmin. They have put a lookup
table in the GPS, based on existing data, which is extrapolated to be valid
up to 2010.

Meindert

 

A fixed table can contain both the current variation plus a "rate of
change per year" term. That's what topo maps do, and it seems to be
accurate enough for the expected life of the map.

Garmin handhelds have internal automatic magnetic variation adjustment,
but they *also* provide manual setting for variation. So if you know
your local variation more accurately than the internal model, you can
set the correct value (to the nearest whole degree).

Dave

 

It's a lookup table built into the firmware.
http://www.edu-observatory.org/maps/maps.html#declination